What Is ROI (Return on Investment)?

ROI stands for Return on Investment: the net profit an investment produces, divided by everything it cost to get there. The formula is as follows.

ROI (%) = (Revenue - Total investment) / Total investment x 100

The part almost every definition skips is what counts as "total investment." It isn't just the money spent on ads. Google Ads' own ROI documentation makes that clear with an example for a physical product: total cost includes "the manufacturing cost of all the items you sold plus your advertising costs," meaning production or product cost plus ad spend, not ad spend on its own.

In practice, for a business selling a physical product, total investment usually adds up: ad spend (SEA, social ads, and similar), product or cost of goods sold, staff time tied to the campaign or to fulfilling those sales, logistics and shipping, and marketing tools or software. A marketing team rarely has all these numbers sitting in one place. Product cost usually lives with purchasing or production, staff cost with HR or internal accounting, and tool costs are split across several departments. That's why a reliable ROI almost always needs figures from finance, not estimates from marketing.

Net profit, the top half of the formula, is what's left once that entire investment is subtracted from the revenue it generated. Subtract only the ad spend, and what you get isn't ROI anymore; at best it's a different flavor of ROAS, and any decision built on it ignores whether the business is actually making money on that campaign.

How wide that cost base gets drawn varies by business. Some teams also allocate a share of fixed costs, like rent or general overhead, while others deliberately stick to costs directly traceable to a given campaign. Which approach gets chosen matters less than staying consistent with it across periods. An ROI calculated with overhead one quarter and without it the next stops being comparable.

One set of numbers is enough to show why ROAS and ROI can say opposite things. A store runs an ad campaign that brings in $4,000 in sales, on $1,000 of ad spend.

Looked at through ROAS alone, the campaign is a success: ROAS = $4,000 / $1,000 = 4, or 4:1. Every dollar put into ads returns $4 in sales. In the ad platform's report, that looks excellent. For the mechanics of how ROAS is calculated and read, see the dedicated ROAS article.

The problem shows up once the rest of the costs those sales actually carry get added in. Producing or buying those products cost $3,000, plus $500 in logistics and order handling: $3,500 on top of the ad spend. Total investment, ads plus everything else, comes to $1,000 + $3,500 = $4,500.

ROI = (4,000 - 4,500) / 4,500 x 100 = -11.1%

With that number, the picture flips completely. The campaign that shone in the ad report with a 4:1 ROAS is actually losing money: it costs more to run than it brings back once the full cost of the business gets counted, not just the ad cost.

This is exactly the mix-up that shows up most often when reading campaign results. A high ROAS speaks to how efficiently the ad budget performed, not to whether the business as a whole makes money on that sale. A campaign can post an enviable ROAS and still run at a loss once the real cost of what was sold gets counted.

ROI and ROAS don't compete with each other: they answer different questions, and both have a place in a well-run marketing account.

ROI answers the question that matters at the business level: is this channel, this product line, or this campaign worth it once everything it actually costs to run gets subtracted? It's the right metric for deciding whether a channel deserves more budget, whether a new product category should launch, or whether a line should be cut despite bringing in sales, because it isn't profitable. Precisely because it needs complete cost data, it's rarely practical to calculate often: at most businesses, ROI gets reviewed monthly or quarterly, once consolidated figures come in from finance.

ROAS fills the gap that cadence leaves open. It can be calculated purely from data already sitting in the ad platform, spend and conversion value, without waiting for accounting to close the month. That makes it ideal for adjusting bids, pausing underperforming ads, or comparing two creatives within the same week. No campaign manager waits for a quarterly close to decide whether to raise or lower a bid.

The healthiest way to use both is in a cascade: ROAS steers the ad account day to day, and ROI periodically checks whether those daily decisions, added up, are building a profitable business. A ROAS that climbs quarter after quarter while ROI stays flat or drops is a clear sign that product cost, logistics, or the rest of the operation is eating into the margin advertising appears to be generating.

In practice, that means ROAS runs as a dashboard metric, often checked daily or weekly, pulled straight from the ad platform. ROI shows up as a calendar item instead, at month-end or after the quarter closes, once finance has delivered the actual cost figures. Teams that try to recalculate ROI weekly usually spend more time chasing down cost numbers than running the campaign itself, without that extra frequency producing better decisions.

• Always pull product, staff, and logistics cost figures from finance before calculating ROI; estimating them from marketing usually inflates the result.
• Use ROAS for tactical day-to-day calls, like adjusting bids, pausing ads, or comparing creatives, and save ROI for budget or channel-continuation decisions.
• Calculate ROI by channel and by product line, not just at the account level; one channel can offset another and hide real losses.
• Review ROI on the same cadence you close out financial figures, usually monthly or quarterly, instead of trying to recalculate it daily.
• Document which cost items feed into each ROI calculation, so the same indicator stays comparable month over month without the criteria shifting underneath it.
• When a campaign's ROAS is high but the ROI doesn't add up, check the product's actual margin before touching the bid.
• Decide once, in writing, whether overhead gets allocated into the ROI calculation, and stick to that decision across every period.

• Calculating ROI using only ad spend as the investment, without adding product, staff, or logistics cost: that produces ROAS dressed up as ROI, not real ROI.
• Using ROI for decisions that need real-time data, like pausing an ad today, when ROAS is better suited for that.
• Comparing ROI across campaigns with very different cost structures, like a physical product with shipping versus a digital service, as if they were directly comparable.
• Calling a campaign good purely because of a high ROAS, without checking the real margin on the product sold.
• Changing what counts as a cost from one month to the next, which makes ROI stop being comparable over time.
• Trying to calculate ROI as often as ROAS, when the cost data it needs simply isn't available at that frequency.